Compositum of Fields

In the world of abstract algebra, fields serve as fundamental universes of numbers with consistent rules for arithmetic. But what happens when we want to merge two of these universes? How do we construct a larger world that contains both, and more importantly, what are its properties? This question lies at the heart of understanding complex algebraic structures. This article tackles this challenge by introducing the compositum of fields, a powerful tool for combining field extensions. We will explore the elegant rules that govern the complexity and symmetry of these composite structures, moving from intuitive analogies to rigorous algebraic principles.

The article is structured to guide you through this concept, beginning with the foundational principles and mechanisms. This first section defines the compositum and unveils the master formula that calculates its degree based on the intersection of the original fields. Following this, the section on applications and interdisciplinary connections will showcase the profound impact of this concept, demonstrating how it is used to unravel the symmetries of polynomials, analyze the arithmetic of number systems, and even solve ancient geometric puzzles. Through this exploration, you will see how the simple act of combining fields provides a unified framework for some of the most beautiful ideas in mathematics.

Imagine you are a linguist studying two ancient languages. Each language has its own unique grammar and vocabulary, a self-contained world of expression. Now, what happens if you discover a third, previously unknown text that seems to blend elements from both? To decipher it, you can't just study one language or the other; you must construct a new, larger framework that encompasses both. You need to build the smallest "super-language" that contains every word and grammatical rule from the originals.

In abstract algebra, we do something very similar with mathematical worlds called ​​fields​​. A field is a set of numbers where you can add, subtract, multiply, and divide without ever leaving the set (except for division by zero, which is, as always, forbidden). The rational numbers, which we call $\mathbb{Q}$, form our base field—the familiar world of fractions.

We can create new, richer worlds by "adjoining" new numbers. For example, by taking all the rational numbers and throwing in $\sqrt{2}$, we create a new field, $\mathbb{Q}(\sqrt{2})$. Its elements look like $a + b\sqrt{2}$, where $a$ and $b$ are rational. This process is called a ​​field extension​​. When we have two such extensions, say $K_1$ and $K_2$, we can combine them to form their ​​compositum​​, written $K_1K_2$. This is the smallest field that contains every number from both $K_1$ and $K_2$, our "super-language" that understands both. The fascinating question is: what does this new, combined world look like? How complex is it?

The first thing we want to know about a new field is its "size" or "complexity" relative to our starting point. We measure this with the ​​degree​​ of the extension, written $[K:F]$. This number isn't about how many elements are in the field (which is usually infinite), but rather about its dimension. Think of it this way: to describe any point on a plane, you need two coordinates (like $x$ and $y$). We say the plane is two-dimensional. Similarly, any number in $\mathbb{Q}(\sqrt{2})$ can be uniquely described with two rational "coordinates," the $a$ and $b$ in $a \cdot 1 + b \cdot \sqrt{2}$. So, we say $[\mathbb{Q}(\sqrt{2}):\mathbb{Q}]=2$.

This leads to a natural and compelling question: if we combine a field $K_1$ of degree $m$ over $\mathbb{Q}$ and a field $K_2$ of degree $n$ over $\mathbb{Q}$, what is the degree of their compositum, $K_1K_2$? The most naive and hopeful guess would be that we simply multiply the complexities: $m \times n$. Sometimes, this is exactly what happens.

Let's explore this simple, beautiful scenario. Suppose we have two extensions that are fundamentally "different" from each other, sharing no common algebraic structure beyond the rational numbers they both grew from.

Consider a hypothetical system where one part is described by a number $\alpha$ which is a root of an irreducible polynomial of degree 5, giving the extension $\mathbb{Q}(\alpha)$ with degree 5. Another part involves a number $\beta$ from an irreducible polynomial of degree 3, giving $\mathbb{Q}(\beta)$ with degree 3. To describe the whole system, we need the compositum $\mathbb{Q}(\alpha, \beta)$. What is its degree? The degrees 5 and 3 are coprime numbers. This means the extensions they represent cannot share any intermediate "sub-extensions" of smaller degree. They are, in an algebraic sense, completely independent. In this ideal situation, our naive guess is correct: the degree of the compositum is the product of the individual degrees. The complexity of the combined system is simply $5 \times 3 = 15$.

This elegant multiplication rule holds whenever the two fields are ​​linearly disjoint​​, meaning their only common ground is the base field $\mathbb{Q}$. This happens, for instance, when we combine $\mathbb{Q}(\sqrt{2})$ and $\mathbb{Q}(\sqrt{3})$. One involves the arithmetic of $\sqrt{2}$, the other $\sqrt{3}$; they don't mix. The degree of their compositum $\mathbb{Q}(\sqrt{2}, \sqrt{3})$ is $2 \times 2 = 4$. Another lovely example involves mixing real and non-real numbers. The field $\mathbb{Q}(\sqrt[3]{2})$ is entirely contained within the real numbers, while the field $\mathbb{Q}(\omega)$, where $\omega = \exp(2\pi i/3)$ is a complex cube root of unity, is not. Their only overlap can be the rational numbers, which are real. So again, their degrees multiply: $[\mathbb{Q}(\sqrt[3]{2}, \omega):\mathbb{Q}] = [\mathbb{Q}(\sqrt[3]{2}):\mathbb{Q}] \cdot [\mathbb{Q}(\omega):\mathbb{Q}] = 3 \times 2 = 6$.

But nature is rarely so simple. What if our two languages are not alien to each other? What if they are more like Spanish and Italian, sharing a significant amount of vocabulary and grammar from their common ancestor, Latin? Combining them doesn't create a world that is twice as complex, because there's a huge overlap.

In field theory, this overlap is the ​​intersection​​ of the two fields, $K_1 \cap K_2$, which is itself a field. The presence of a non-trivial intersection reduces the complexity of the combined world. The wonderfully precise relationship is captured by a master formula: $[K_1K_2 : F] = \frac{[K_1:F][K_2:F]}{[K_1 \cap K_2 : F]}$ This formula is one of the most satisfying in field theory. It tells us that the degree of the compositum is the product of the individual degrees, corrected by—divided by—the degree of their shared part. The larger the overlap, the smaller the compositum.

Let's see this principle in action. Suppose we have an extension $K_1$ of degree 6 and another $K_2$ of degree 10. The degree of their intersection, $[K_1 \cap K_2:\mathbb{Q}]$, must be a number that divides both 6 and 10. The common divisors of 6 and 10 are just 1 and 2.

• If the intersection is just $\mathbb{Q}$, its degree is 1. The compositum degree is $\frac{6 \times 10}{1} = 60$.
• If the intersection is a quadratic field (degree 2), the compositum degree is $\frac{6 \times 10}{2} = 30$. Without more information, we can't know which it is, but we know the only possibilities are 30 and 60.

Let's make this more concrete. When is the degree of $\mathbb{Q}(\sqrt{d_1}, \sqrt{d_2})$ not 4? When does it collapse to 2? According to our formula, this happens when the denominator, the degree of the intersection, is 2, not 1. This means the intersection $\mathbb{Q}(\sqrt{d_1}) \cap \mathbb{Q}(\sqrt{d_2})$ is a quadratic field. But since both are quadratic fields, this can only happen if they are the exact same field. For example, $\mathbb{Q}(\sqrt{12})$ and $\mathbb{Q}(\sqrt{3})$ are in fact the same field, because $\sqrt{12} = 2\sqrt{3}$, and the factor of 2 doesn't change the field itself. So, $[\mathbb{Q}(\sqrt{12}, \sqrt{3}):\mathbb{Q}] = [\mathbb{Q}(\sqrt{3}):\mathbb{Q}] = 2$. The intersection was the entire field itself, and the degree of the compositum collapsed. The "overlap" was total.

Sometimes the overlap is hidden. Consider the fields $K_1 = \mathbb{Q}(\sqrt[3]{2})$ and $K_2 = \mathbb{Q}(\zeta_3\sqrt[3]{2})$, where $\zeta_3$ is a complex cube root of unity. Both have degree 3 over $\mathbb{Q}$. If their intersection were trivial, the compositum would have degree 9. But notice what we can do inside the compositum $K_1K_2$: we can take the element $\zeta_3\sqrt[3]{2}$ from $K_2$ and divide it by the element $\sqrt[3]{2}$ from $K_1$. The result, $\zeta_3$, must also be in the compositum. This reveals that $K_1K_2$ actually contains all the elements of $\mathbb{Q}(\sqrt[3]{2})$ and also $\zeta_3$. So, $K_1K_2$ is the same field as $\mathbb{Q}(\sqrt[3]{2}, \zeta_3)$, which we already saw has degree 6. A naive application of the master formula, $6 = \frac{3 \times 3}{[K_1 \cap K_2 : \mathbb{Q}]}$, would imply an intersection degree of 1.5, which is impossible. What went wrong? The formula is only guaranteed to hold when at least one of the extensions is a Galois extension, a condition not met here. In fact, the intersection $K_1 \cap K_2$ is just $\mathbb{Q}$ (degree 1). The compositum is the larger field $\mathbb{Q}(\sqrt[3]{2}, \zeta_3)$ itself. This example shows how building the compositum can reveal hidden algebraic structures, but also serves as a caution to always check the preconditions of a theorem.

The degree tells us about complexity, but there's a deeper, more beautiful story about structure and symmetry. For certain "nice" extensions, called ​​Galois extensions​​, the field's internal structure is perfectly mirrored by the structure of a finite group of its symmetries, the ​​Galois group​​. How does combining fields affect their symmetries?

A remarkable fact is that the compositum of Galois extensions is itself a Galois extension. For example, the fields $K_1 = \mathbb{Q}(\sqrt{2})$ and $K_2 = \mathbb{Q}(\sqrt{-3})$ are both splitting fields (the fields you need to contain all roots of a polynomial), which are prime examples of Galois extensions. Their compositum, $K_1K_2 = \mathbb{Q}(\sqrt{2}, \sqrt{-3})$, is also a Galois extension. It can even emerge in surprising places. The polynomial $p(x) = x^4 + 2x^2 + 25$, which looks unrelated, actually has all four of its roots inside this very compositum field. In fact, the compositum is precisely the splitting field for $p(x)$, and so its degree is 4. This demonstrates a profound unity: the world built by combining the solutions to $x^2-2=0$ and $x^2+3=0$ is the very same world needed to solve $x^4+2x^2+25=0$.

The Galois group of a compositum holds the key to its structure.

• If one field is contained within another, like $K_q \subseteq K_p$, their compositum is simply the larger field, $K_p$. The problem of finding the Galois group of $K_p K_q$ just becomes finding the group of $K_p$. The "combination" was really an absorption.
• If the fields are disjoint (intersecting only at $\mathbb{Q}$), their symmetries combine in the simplest way possible: the Galois group of the compositum is the ​​direct product​​ of the individual Galois groups, $\text{Gal}(K_1K_2/\mathbb{Q}) \cong \text{Gal}(K_1/\mathbb{Q}) \times \text{Gal}(K_2/\mathbb{Q})$. The symmetries of one world act independently of the symmetries of the other.

But what about the general case, where there's a non-trivial overlap? This is where the true power of Galois theory shines. There's a stunning isomorphism that connects the structure of the compositum to the intersection: $\text{Gal}(LF/F) \cong \text{Gal}(L/(L \cap F))$ Let's unpack this. We have two extensions, $L/K$ and $F/K$. The left side is the Galois group of the compositum $LF$ viewed from the perspective of the field $F$. It describes the symmetries of the combined system that leave all of $F$ unchanged. The right side is the Galois group of $L$ over the intersection $L \cap F$. It describes the symmetries of $L$ that leave the shared part, $L \cap F$, unchanged. The theorem says these two groups are identical! Symmetries of the whole that fix one part correspond to symmetries of the other part that fix the overlap.

Consider $L = \mathbb{Q}(\sqrt{3}, \sqrt{5})$ and $F = \mathbb{Q}(\sqrt{5}, \sqrt[3]{2})$. Their intersection is clearly seen to contain $\mathbb{Q}(\sqrt{5})$. Using degree arguments, we can show this is the entire intersection: $L \cap F = \mathbb{Q}(\sqrt{5})$. Our theorem then tells us that $\text{Gal}(LF/F)$ is isomorphic to $\text{Gal}(L/\mathbb{Q}(\sqrt{5}))$. The Galois group of $L/\mathbb{Q}$ consists of four symmetries, but only those that fix $\sqrt{5}$ will be in $\text{Gal}(L/\mathbb{Q}(\sqrt{5}))$. This is just the automorphism that sends $\sqrt{3} \to -\sqrt{3}$ and its identity. This is a group of order 2. The complex symmetry group of the compositum is reduced to a simple subgroup determined entirely by the geometry of the intersection.

This connection is so tight that we can reason about fields just by studying groups. For a Galois extension whose group is the quaternion group $Q_8$, we know from group theory that it has exactly three subgroups of order 4. By the Galois correspondence, this means the extension has exactly three quadratic subfields. The intersection of any two of these distinct order-4 subgroups is always the center of $Q_8$, which has order 2. The compositum of the corresponding fields must then correspond to this intersection of subgroups. The degree of the compositum field is therefore the index of this intersection subgroup, which is $|Q_8|/2 = 8/2=4$. We can deduce that combining any two distinct quadratic subfields of a $Q_8$ extension always produces a biquadratic extension of degree 4, a conclusion reached from pure group theory, without touching a single polynomial.

This journey, from adding numbers to uncovering interlocking symmetries, reveals the essence of modern algebra. The structure of combined fields is not arbitrary. It is governed by elegant and powerful principles that connect the size of an extension, the extent of its overlap with another, and the very nature of its symmetries. The master formula for degrees and the isomorphism for Galois groups are two sides of the same coin, a beautiful testament to the hidden unity in the abstract world of fields.

Having understood the principles and mechanisms of the compositum of fields, you might be asking yourself, "What is it all for?" It is a fair question. Abstract algebra can sometimes feel like a game played with symbols, detached from reality. But here, with the concept of the compositum, we hold a key—a master tool for construction. It's like learning the principle of the arch in architecture. Suddenly, you can build bridges and cathedrals. The compositum allows us to build complex mathematical structures from simpler ones and, most wonderfully, to understand the properties of the grand structure by studying its constituent parts. Let us embark on a journey to see how this single idea builds bridges across different landscapes of mathematics and beyond.

At its heart, Galois theory is the study of symmetries in the solutions of polynomial equations. The symmetries of a given polynomial's roots are captured by a group, its Galois group. But what happens if we have a polynomial that is the product of two simpler ones, say $h(x) = f(x)g(x)$? The roots of $h(x)$ are simply the roots of $f(x)$ together with the roots of $g(x)$. The "world" containing all these roots, the splitting field of $h(x)$, is precisely the compositum of the splitting fields of $f(x)$ and $g(x)$.

Now for the magic. Imagine the polynomials $f(x) = x^3 - 2$ and $g(x) = x^2 - 5$. The first has a set of symmetries isomorphic to $S_3$, the group of permutations of three objects. The second has a much simpler symmetry group, $C_2$, which just swaps $\sqrt{5}$ and $-\sqrt{5}$. What is the symmetry group of their combination? One might fear a terribly complicated new group. But it turns out that the fields generated by their roots, $\mathbb{Q}(\sqrt[3]{2}, \omega)$ and $\mathbb{Q}(\sqrt{5})$, are "linearly disjoint"—they only share the rational numbers $\mathbb{Q}$ that we started with. Because they do not interfere with each other, their symmetries combine in the simplest way imaginable: the Galois group of the compositum is the direct product of the individual groups, $S_3 \times C_2$. This is a profound principle: when worlds are independent, their symmetries compose directly. The abstract rule underpinning this is that the process of building a splitting field over a larger base field is compatible with forming the compositum first.

Number theory is rife with special number systems, or fields, that have beautiful properties. Among the most important are the cyclotomic fields, $\mathbb{Q}(\zeta_n)$, which are formed by adjoining a primitive $n$-th root of unity to the rational numbers. What happens when we combine two such fields, say $\mathbb{Q}(\zeta_6)$ and $\mathbb{Q}(\zeta_{10})$? Do we get some strange new object? The answer is astonishingly elegant. The compositum of $\mathbb{Q}(\zeta_m)$ and $\mathbb{Q}(\zeta_n)$ is simply $\mathbb{Q}(\zeta_l)$, where $l$ is the least common multiple of $m$ and $n$. The abstract algebraic operation of forming a compositum corresponds to the elementary arithmetic operation of finding the lcm!

This principle allows us to construct and analyze abelian Galois groups with remarkable ease. If we take the compositum of cyclotomic fields whose indices are powers of distinct primes, such as $\mathbb{Q}(\zeta_{27})$, $\mathbb{Q}(\zeta_{25})$, and $\mathbb{Q}(\zeta_{16})$, their respective Galois groups combine as a direct product. This allows us to predict the precise structure of the resulting Galois group, breaking it down into its elementary cyclic components.

The power of the compositum in arithmetic goes even deeper. Consider the biquadratic field $K = \mathbb{Q}(\sqrt{5}, \sqrt{13})$, which is the compositum of two simpler quadratic fields. A central question in number theory is how prime numbers like 3, 5, 7, 11... behave when viewed within a larger number field. Do they remain prime, or do they split into products of new primes? For our field $K$, the answer depends beautifully on the parts from which it was built. A prime $p$ splits completely in $K$ if and only if it splits in both $\mathbb{Q}(\sqrt{5})$ and $\mathbb{Q}(\sqrt{13})$. This translates, via the law of quadratic reciprocity, to simple conditions on $p$ modulo 5 and modulo 13. The compositum structure provides a perfect dictionary between the abstract structure of the Galois group and the concrete splitting behavior of primes.

The compositum of fields has a surprising connection to a problem that fascinated the ancient Greeks: which geometric lengths can be constructed using only a compass and an unmarked straightedge? The modern algebraic answer is that a number is constructible only if the degree of the field it generates over $\mathbb{Q}$ is a power of 2.

Now, suppose we have two constructible numbers, $\alpha$ and $\beta$. What about numbers formed from them, like $\alpha + \beta$ or $\alpha \beta$? These numbers live in the compositum field $\mathbb{Q}(\alpha, \beta)$. The degree of this composite field is given by the formula $[ \mathbb{Q}(\alpha, \beta) : \mathbb{Q} ] = \frac{[\mathbb{Q}(\alpha):\mathbb{Q}][\mathbb{Q}(\beta):\mathbb{Q}]}{[\mathbb{Q}(\alpha)\cap\mathbb{Q}(\beta):\mathbb{Q}]}$. Since $\alpha$ and $\beta$ are constructible, the degrees in the numerator are powers of 2. The intersection is also a subfield, so its degree is also a power of 2. The result? The degree of the compositum is guaranteed to be a power of 2. This shows that the set of all constructible numbers is itself a field—a world closed under its own arithmetic, built piece by piece through composita of quadratic extensions.

We can take this idea to its logical extreme. Let's imagine building an enormous field, let's call it $\mathcal{F}_2$, by forming the compositum of all possible finite real Galois extensions of $\mathbb{Q}$ whose Galois groups are 2-groups. This field contains the constructible numbers and infinitely more. It is a vast algebraic universe. But is it everything? Could a number like $\sqrt{\pi}$ exist in this world? The answer reveals a fundamental truth about the nature of numbers. By its very construction, every single element in the colossal field $\mathcal{F}_2$ must be algebraic—it must be a root of some polynomial with rational coefficients. However, Lindemann proved in 1882 that $\pi$ is transcendental, and it follows that $\sqrt{\pi}$ is also transcendental. Therefore, $\sqrt{\pi}$ cannot be an element of $\mathcal{F}_2$. The compositum, for all its power to build infinite algebraic structures, cannot bridge the chasm between the algebraic and the transcendental.

In the highest reaches of modern number theory, the compositum is not just a tool; it is the language used to describe the very architecture of our number systems.

Every number field has a fundamental invariant called its discriminant, a single number that encodes deep information about its arithmetic structure. If we build a composite field $K = K_1K_2$ from two fields $K_1$ and $K_2$ with coprime discriminants, the discriminant of the new field is given by the breathtakingly simple formula $d_K = d_{K_1}^{n_2} d_{K_2}^{n_1}$, where $n_1$ and $n_2$ are the degrees of the component fields. The "DNA" of the composite structure is determined in a precise and elegant way by the DNA of its parts.

Perhaps the most profound application is in local class field theory, a monumental achievement of 20th-century mathematics. This theory provides a complete description of all abelian extensions of a local field $K$ (a completion of a number field, like the $p$-adic numbers). The central object, the maximal abelian extension $K^{\text{ab}}$, seems unimaginably complex. Yet, the theory reveals that it has a stunningly simple architecture: $K^{\text{ab}}$ is the compositum of two simpler, "purer" extensions.

1. The maximal unramified extension, $K^{\text{ur}}$.
2. A totally ramified extension, known as a Lubin-Tate extension, $K_{\varpi}^{\text{LT}}$.

These two fields are linearly disjoint, their only intersection being $K$ itself. This means that every abelian extension of $K$ can be understood as being built from these two fundamental, and in a sense orthogonal, types of extensions. The compositum concept provides the framework for this grand unification, showing how the most intricate structures can be resolved into a composite of elemental parts. It is a testament to the power of a simple idea to bring clarity, beauty, and profound understanding to the deepest questions of mathematics.